Synthesis of the Isoquinocycline− Pyrrolopyrrole Substructure

Aug 10, 2010 - Fachbereich Chemie, Philipps-University Marburg, Hans-Meerwein-Strasse, D-35032 Marburg, Germany. Org. Lett. , 2010, 12 (17), pp 3808â€...
0 downloads 0 Views 985KB Size
ORGANIC LETTERS

Synthesis of the Isoquinocycline-Pyrrolopyrrole Substructure

2010 Vol. 12, No. 17 3808-3811

Jens Cordes, Klaus Harms, and Ulrich Koert* Fachbereich Chemie, Philipps-UniVersity Marburg, Hans-Meerwein-Strasse, D-35032 Marburg, Germany [email protected] Received June 30, 2010

ABSTRACT

The synthesis of the pyrrolopyrrole substructure of the isoquinocyclines is reported. The pentacyclic (CDEFG) substructure of isoquinocycline A and B, which contains an unusual 2,4,5,6-tetrahydropyrrolo[2,3-b]pyrrole (FG) connected via an N,O-spiro acetal to the anthraquinoid core of the isoquinocycline aglycon has been synthesized. Key steps were a nickel(0)-mediated hydrocyanation of an alkynone, the conversion of an O,O-acetal into an N,O-acetal, and an intramolecular amidine alkylation.

Quinocyclines1 are a class of anthracycline related natural products2 consisting of the four compounds quinocycline A, quinocycline B, isoquinocycline A, and isoquinocycline B (Figure 1). All four substances were first isolated in the 1950s from Streptomyces aureofaciens.1 Quinocycline B, also known as kosinostatin,3 was also isolated from Micromono(1) (a) Celmer, W. D.; Murai, K.; Rao, K. V.; Tanner, F. W.; Marsh, W. S. Antibiotics Annual 1957-1958; Medical Encyclopedia Inc.: New York, 1958; pp 484-492. (b) McBride, T. J.; English, A. R. Antibiotics Annual 1957-1958; Medical Encyclopedia Inc.: New York, 1955; pp 493-501. (c) Martin, J. H.; Shay, A. J.; Pruess, L. M.; Porter, J. N.; Mowat, J. H.; Bohonos, N. Antibiotics Annual 1954-1955; Medical Encyclopedia Inc.: New York, 1955; pp 1020-1024. (d) Cosulich, D. B.; Mowat, J. H.; Broschard, R. W.; Patrick, J. B.; Meyer, W. E. Tetrahedron Lett. 1963, 7, 453–458. Erratum: Tetrahedron Lett. 1964, 13, 750. (e) Tulinsky, A. J. Am. Chem. Soc. 1964, 86, 5368–5369. (2) For a review on naturally occurring anthracyclines, see: Laatsch, H.; Fotso, S. Top. Curr. Chem. 2008, 282, 3–74. (3) (a) Furumai, T.; Igarashi, Y.; Higuchi, H.; Saito, N.; Oki, T. J. Antibiot. 2002, 55, 128–133. Corrections: J. Antibiot. 2003, 56, C1. (b) Igarashi, Y.; Higuchi, H.; Oki, T.; Furumai, T. J. Antibiot. 2002, 55, 134– 140. Corrections: J. Antibiot. 2003, 56, C1. 10.1021/ol101500k  2010 American Chemical Society Published on Web 08/10/2010

Figure 1. Quinocyclines.

spora sp. (TP-A0468, along with isoquinocycline B)3 and Streptomyces Violaceusniger (HAL64).4 Apart from their antibiotic and cytotoxic activities, some unusual structural characteristics of the quinocyclines make (4) El-Naggar, M. Y. J. Microbiol. 2007, 45, 262–267.

them interesting synthetic targets. All quinocyclines show an anthraquinoid tetracycle (ABCD) and are glycosylated at C10 with γ-branched octoses. These are trioxacarcinose B for (iso)quinocycline B and dihydro-trioxacarcinose B for (iso)quinocycline A.5 The most remarkable substructure of the (iso)quinocycline aglycon is the bicyclic amidine (FG) that connects C7 and C9O via an N,O-spiro center. This heterocycle, a 2,4,5,6tetrahydropyrrolo[2,3-b]pyrrole, is unique among natural products and has so far only been found in the quinocyclines. Having completed the stereoselective synthesis of dihydrotrioxacarcinose B,6 we report here a synthetic access to the pyrrolopyrrole substructure as the next step toward the total synthesis of the quinocyclines. A retrosynthetic analysis (Scheme 1) of the CDEFGsubstructure 2 of the (iso)quinocycline aglycon (1) reveals

Scheme 2. Synthesis of the Lactone 4

Scheme 1. Retrosynthetic Analysis of the CDEFG-Substructure of the (Iso)quinocycline Aglycon

the possibility to construct the pyrrolopyrrole (FG) from the (Z)-β-nitrilo alkenone 3. By converting the nitrile to an amidine and the alcohol to a leaving group, the G-ring could be closed by intramolecular substitution, while the F-ring is established by formation of an N,O-acetal. Enone 3 should be accessible by a cyanide addition to the corresponding alkynone, which originates from lactone 4. The diol motif in 4 leads to the olefin precursor 5, which could be prepared from the arylacetic acid 6. This results in a synthetic strategy that starts with the ring C and adds all other rings sequentially. This approach should be applicable to the isoquinocycline and quinocycline series. The synthesis of the racemic lactone 4 is outlined in Scheme 2. The starting point was the conversion of methacrolein (7) to the iodo dioxolane 8.7 Fischer esterification of 6 afforded the ethyl ester 9, which was alkylated with the iodide 8 to obtain the dioxolane 10. Friedel-Crafts cyclization and subsequent elimination resulted in the formation of (5) (a) Matern, U.; Grisebach, H.; Karl, W.; Achenbach, H. Eur. J. Biochem. 1972, 29, 1–4. (b) Matern, U.; Grisebach, H. Eur. J. Biochem. 1972, 29, 5–11. (6) Ko¨nig, C. M.; Harms, K.; Koert, U. Org. Lett. 2007, 9, 4777–4779. (7) Larson, G. L.; Klesse, R. J. Org. Chem. 1985, 50, 3627. Org. Lett., Vol. 12, No. 17, 2010

the dihydronaphthalene 11. The ethyl ester could be hydrolyzed to the γ,δ-unsaturated carboxylic acid 5, which was suitable for an iodolactonization. The resulting iodolactone proved to be unstable and was therefore converted directly with potassium methoxide into the epoxide 12. Under aqueous acidic conditions, the epoxide was opened regioselectively at the benzylic position, yielding a trans-diol that underwent a subsequent lactonization to form the hydroxylactone 13. The configuration of this compound could be confirmed by an X-ray crystal structure (14). The TBSprotection of the alcohol finally lead to the lactone 4. It turned out that the reaction sequence from 11 to 4 works quite well without intermediate chromatography. As compounds 9, 10, and 11 can be purified by distillation, the TBSether 4 can be prepared on a gram scale with only one final chromatographic purification step. The next steps of the synthesis are shown in Scheme 3. The lactone 4 was allowed to react with double lithiated 3-butyn-1-ol to give the alkynone 15. No hemiacetal formation was observed for the case of the hydroxy alkynone 15. The hydrocyanation of this alkynone was conducted according to a method developed by Arzoumanian et al.8 By generating [Ni0(CN)4]4- from a nickel precursor, potassium cyanide, and zinc as a reducing agent, Arzoumanian et al. were able to convert non-4-yn-3-one to 3-butyl-5-ethyl-5hydroxy-1H-pyrrol-2(5H)-one in 65% yield.8 By applying these conditions to our alkynone 15, we therefore expected to get the hydroxypyrrolone 17 or its condensation product (8) Arzoumanian, H.; Jean, M.; Nuel, D.; Garcia, J. L.; Rosas, N. Organometallics 1997, 16, 2726–2729. 3809

Scheme 3. Nickel(0)-Mediated Hydrocyanation Leading to the Formation of the Cyclic Imidate 16

compared to the proposed mechanism for the formation of pyrrolones, which begins with a partial hydrolysis of the nitrile group by water.8 While compound 16 contained all carbon atoms of the target already, two problems still remained to be solved: the conversion of the O,O-acetal into an N,O-acetal and the closure of the G ring. N- and O-tosylation of compound 16 gave compound 20 in very good yield (Scheme 5). The tosylated imidate 20

Scheme 5. Synthesis of the Amidines 21 and 22

18 each with an N,O-acetal. Instead, the cyclic imidate 16 with an O,O-acetal was formed. This result, the formation of a cyclic imidate via a β-nitrilocarbonyl intermediate instead of the formation of a pyrrolone, is in contrast to almost all examples in the literature.8,9 There is only one example for the formation of a benzannelated cyclic imidate starting from 2-cyano-benzaldehyde.10 The structural assignment of the imidate formation was confirmed by an X-ray crystal structure (19) which was obtained from the TMS-analogue of TBS-ether 16. A mechanistic explanation for the formation of the unusual imidate is possible from considering the structural context of the intermediate β-nitrilo-ketone 3. The tertiary alcohol of the β-nitrilo-ketone 3 (Scheme 4) is in close proximity to

Scheme 4. Plausible Mechanism for the Formation of the Imidate 16 from Alkynone 15

the ketone, which allows easy formation of a hemi acetal via a five-membered ring. On the other hand, the ketone is already preorientated for an attack on the nitrile to form a second five-membered ring. These two ring formations may occur in concerted or sequential fashion; in any case, the intramolecular attack of the OH group should be favored 3810

was easily attacked by ammonia. If this addition was carried out in the presence of TMSCl, the two N,O-acetals 21 and 22 could be obtained. Throughout the synthesis so far, only one single diastereomer was obtained for the spiro center at the O,O-acetal. At this point, two epimeric N,O-acetals 21 and 22 were formed. Compound 21 corresponds to the configuration of isoquinocycline A/B, while compound 22 exhibits the configuration of quinocycline A/B. The two epimers could be separated by chromatography. A slow epimerization of quinocycline into the more stable isoquinocycline was reported for the natural product.1,3 No epimerization at the N,O-acetal was observed during the purification and storage of compounds 21 and 22. The completion of the synthesis of amidine 2 (the CDEFGsubstructure of the aglycon of isoquinocycline A and B) is outlined in Scheme 6. The desired ring closure to the pyrrolopyrrole substructure via an intramolecular amidine alkylation was achieved by the treatment of the amidine 21 with lutidine at 120 °C for several hours. The structural assignment of the resulting pyrrolopyrrole 23 was confirmed by an X-ray crystal structure (24). The deprotection of 23 (9) (a) Rosas, N.; Cabrera, A.; Sharma, P.; Arias, J. L.; Garcı´a, J. L.; Arzoumanian, H. J. Mol. Catal. A: Chem. 2000, 156, 103–112. (b) Zanatta, N.; Rosa, L. S.; Cortelini, M. F. M.; Beux, S.; Santos, A. P. D.; Bonacorso, H. G.; Martins, M. A. P. Synthesis 2002, 16, 2404–2408. (c) Nenajdenko, V. G.; Druzhinin, S. V.; Balenkova, E. S. Russ. Chem. Bull., Int. Ed. 2003, 52, 2467–2472. (d) Martins, M. A. P.; Pereira, C. M. P.; Zimmermann, N. E. K.; Cunico, W.; Moura, S.; Beck, P.; Zanatta, N.; Bonacorso, H. G. J. Fluorine Chem. 2003, 123, 261–265. (10) Sato, R.; Ohmori, M.; Kaitani, F.; Kurosawa, A.; Senzaki, T.; Goto, T.; Saito, M. Bull. Chem. Soc. Jpn. 1988, 61, 2481–2485. Org. Lett., Vol. 12, No. 17, 2010

Scheme 6. Completion of the Synthesis of Amidine 2, the CDEFG-Substructure of Isoquinocycline A and B

with TBAF gave alcohol 25, which was further deprotected with HF-triethylamine to deliver the free amidine 2. In contrast to the smooth conversion of amidine 21 to the CDEFG-substructure of the isoquinocyclines, the analogue conversion of amidine 22 to the CDEFG-quinocycline substructure proved to be problematic. Under the same conditions that allowed the synthesis of pyrrolopyrrole 23 from amidine 21, the epimeric amidine 22 gave the lactam 27 as the main product. The isolated compound 27 represents the hydrolysis product of the desired pyrrolopyrrole 26 (Scheme 7). One possible explanation for this failure could be that, after successful ring closure, the less stable Ntosylated amidine with the quinocycline configuration hydrolyzed easily during workup and purification. Optimal

Org. Lett., Vol. 12, No. 17, 2010

Scheme 7. Attempted Synthesis of the Pyrrolopyrrole 26

cyclization conditions for the intramolecular amidine alkylation in the quinocycline series remain to be elaborated. Future work will be directed at the extension of the present synthesis to the complete heptacyclic aglycon of the (iso)quinocyclines and its application to a stereoselective total synthesis of (iso)quinocycline A and B. In conclusion, an efficient synthesis for the CDEFGsubstructure of isoquinocycline A and isoquinocycline B was developed. Key steps were a nickel(0)-mediated hydrocyanation of an alkynone to install the carbon skeleton, the conversion of a cyclic imidate to an amidine (O,O- to N,Oacetal), and the intramolecular amidine alkylation for the final G-ring closure. A synthetic solution for the long-standing problem of 2,4,5,6-tetrahydropyrrolo[2,3-b]pyrrole, which is connected via an N,O-spiro acetal to the anthraquinoid core of the quinocyclines, has been found. Acknowledgment. Generous support by the Deutsche Forschungsgemeinschaft (KO1349/8-2) and the Fonds der Chemischen Industrie (Doktorandenstipendium for J.C.) is gratefully acknowledged. Supporting Information Available: Experimental procedures and analytical data for all new compounds and single X-ray crystallographic data (CIF) for 14, 19, and 24. This material is available free of charge via the Internet at http://pubs.acs.org. OL101500K

3811